You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Simplify and refactor element classes for nested Matrices/Vectors #36
Framework in its current state defines, from scratch, home-grown Vector2 and Matrix2 classes (implementations based on those in InfiniTAM).
This is unnecessary and difficult to expand to alternative element types, i.e. Vector3, Matrix4, etc. Instead, these should be just classes extending Eigen's own Vector3, Matrix2, etc., adding the missing functionality.
For instance, Vector2f_uv class could add a Vector2f(single_value) constructor, u() and v() member functions (getters), etc.
This might not be such a good idea -- the homegrown Vector2/Vector3/Matrix2 etc. classes are CUDA-compatible, whereas I'm not sure that the resulting Eigen classes will be.
Activity
Algomorph commentedon Mar 4, 2019
This might not be such a good idea -- the homegrown Vector2/Vector3/Matrix2 etc. classes are CUDA-compatible, whereas I'm not sure that the resulting Eigen classes will be.